Packet 7 (KW first pass)

Last updated: 2019-04-13

This is Kara’s first pass at looking at the Packet 7 data (with data from 2019-03-18). (Note: PDF printed portrait, 80% zoom, minimum margins.)

Confirmatory hypotheses (as specified in preregistration)

Note: I included all participants, even if they failed attention checks, since we did not specify that we would exclude participants in the preregistration. At some point we should try running everything on the subset of participants who passed all attention checks just so that we know, but this decreases our sample sizes pretty dramatically.

Hypothesis #1

Hypothesis #1: People who score higher on the Tellegen Absorption Scale (Tellegen & Atkinson, 1974) also report more spiritual experiences.

Preregistered analysis: “We will standardize scores on each measure, collapsing across field sites, and then conduct a mixed-effects linear regression, using absorption scores to predict reports of spiritual experiences, allowing for differences across our five field sites and across the two scales we used to measure spiritual experience. Using the “lme4” package for R (Bates, 2018), we will specify this model as follows: experience ~ absorption * site * scale + (1 | subject), where “scale” is a categorical variable comparing the two spiritual experience scales (effect-coded), and “site” is a categorical variable comparing the five field sites (effect-coded). Our prediction is that the main effect of absorption on experience will be significantly positive (b > 0, t > 2, p < 0.05 using Satterthwaite’s degrees of freedom method as implemented in the “lmerTest” package for R, Kuznetsova, Brockhoff, & Christensen, 2017). (The rest of the parameters included in the model will be of great interest to us, but are exploratory.)

If we have trouble fitting these models or there are other signs of model fit problems, we will consider the following alternative model: experience ~ absorption * scale + (1 | site/subject)."

Here are our results:

parameter	b	standard error	df	t	p	significant
(Intercept)	-0.07	0.03	594	-2.36	0.019	*
abs	0.27	0.03	594	8.66	<0.001	*
p7_ctry1	-0.32	0.06	594	-5.55	<0.001	*
p7_ctry2	0.79	0.05	594	16.70	<0.001	*
p7_ctry3	-0.13	0.06	594	-2.26	0.024	*
p7_ctry4	-0.91	0.07	594	-13.74	<0.001	*
scale1	0.00	0.02	594	0.15	0.878
abs:p7_ctry1	-0.02	0.05	594	-0.43	0.669
abs:p7_ctry2	0.23	0.05	594	4.93	<0.001	*
abs:p7_ctry3	0.06	0.07	594	0.89	0.372
abs:p7_ctry4	-0.10	0.08	594	-1.35	0.178
abs:scale1	-0.02	0.02	594	-0.86	0.391
p7_ctry1:scale1	-0.04	0.03	594	-1.16	0.246
p7_ctry2:scale1	-0.05	0.03	594	-1.84	0.066
p7_ctry3:scale1	0.10	0.04	594	2.82	0.005	*
p7_ctry4:scale1	-0.20	0.04	594	-5.06	<0.001	*
abs:p7_ctry1:scale1	-0.02	0.03	594	-0.72	0.47
abs:p7_ctry2:scale1	-0.03	0.03	594	-1.24	0.215
abs:p7_ctry3:scale1	0.01	0.04	594	0.23	0.821
abs:p7_ctry4:scale1	0.04	0.05	594	0.80	0.427

We found evidence in favor of Hypothesis #1: As predicted, the main effect of absorption was significantly positive (see bolded row in table above).

For every increase of 1 standard deviation in absorption, participants (on average) increased by 0.27 standard deviations in their spiritual experiences. In concrete terms, this means that for every increase of 7.24 points on the Absorption scale (range: 0-34), participants (on average) increased by 4.56 points on the Daily Spiritual Experiences scale (range: 0-70) or 4.79 points on the Spiritual Events scale (range: 0-80).

Hypothesis #2

Hypothesis #2: People who score higher on our two porosity indices also report more spiritual experiences.

Preregistered analysis: “We will standardize scores on each measure, collapsing across field sites, and then conduct a mixed-effects linear regression, using porosity scores (from each of our two indices) to predict reports of spiritual experiences across the 2 spiritual experience scales included in our survey packet, allowing for differences across our five field sites and across the two scales we used to measure spiritual experience. Using the “lme4” package for R (Bates, 2018), we will specify this model as follows: experience ~ porosity * site * scale + (1 | subject), where “scale” is a categorical variable comparing the two spiritual experience scales (effect-coded), and “site” is a categorical variable comparing the five field sites (effect-coded). We will run two versions of this model, one using each of the two “porosity” measures described above. For each regression, our prediction is that the main effect of porosity on experience will be significantly positive (b > 0, t > 2, p < 0.05 using Satterthwaite’s degrees of freedom method as implemented in the “lmerTest” package for R, Kuznetsova, Brockhoff, & Christensen, 2017). (The rest of the parameters included in the model will be of great interest to us, but are exploratory.)

If we have trouble fitting these models or there are other signs of model fit problems, we will consider the following alternative model: experience ~ porosity * scale + (1 | site/subject)."

Here are our results, using the Porosity scale (Dulin):

parameter	b	standard error	df	t	p	significant
(Intercept)	-0.04	0.04	594	-1.13	0.258
por	0.43	0.04	594	12.00	<0.001	*
p7_ctry1	-0.12	0.06	594	-1.94	0.052
p7_ctry2	0.32	0.05	594	6.27	<0.001	*
p7_ctry3	-0.10	0.06	594	-1.56	0.119
p7_ctry4	-0.51	0.08	594	-6.14	<0.001	*
scale1	0.02	0.02	594	0.99	0.324
por:p7_ctry1	0.12	0.06	594	2.03	0.043	*
por:p7_ctry2	0.18	0.05	594	3.66	<0.001	*
por:p7_ctry3	-0.06	0.07	594	-0.89	0.373
por:p7_ctry4	-0.02	0.08	594	-0.25	0.802
por:scale1	0.02	0.02	594	0.92	0.356
p7_ctry1:scale1	-0.06	0.04	594	-1.54	0.124
p7_ctry2:scale1	0.01	0.03	594	0.26	0.794
p7_ctry3:scale1	0.09	0.04	594	2.26	0.024	*
p7_ctry4:scale1	-0.12	0.05	594	-2.27	0.024	*
por:p7_ctry1:scale1	-0.01	0.04	594	-0.20	0.842
por:p7_ctry2:scale1	-0.17	0.03	594	-5.23	<0.001	*
por:p7_ctry3:scale1	0.00	0.05	594	0.01	0.989
por:p7_ctry4:scale1	0.09	0.05	594	1.76	0.079

And here are our results, using the Martha Story scale:

parameter	b	standard error	df	t	p	significant
(Intercept)	-0.05	0.03	594	-1.40	0.163
mm	0.32	0.04	594	8.76	<0.001	*
p7_ctry1	-0.07	0.06	594	-1.15	0.252
p7_ctry2	0.33	0.05	594	6.42	<0.001	*
p7_ctry3	-0.17	0.06	594	-2.86	0.004	*
p7_ctry4	-0.64	0.08	594	-8.07	<0.001	*
scale1	0.02	0.02	594	1.28	0.202
mm:p7_ctry1	0.15	0.06	594	2.38	0.018	*
mm:p7_ctry2	0.14	0.05	594	2.90	0.004	*
mm:p7_ctry3	-0.06	0.07	594	-0.92	0.356
mm:p7_ctry4	0.01	0.09	594	0.06	0.95
mm:scale1	-0.04	0.02	594	-1.78	0.076
p7_ctry1:scale1	-0.11	0.04	594	-2.85	0.005	*
p7_ctry2:scale1	0.04	0.03	594	1.45	0.148
p7_ctry3:scale1	0.07	0.03	594	2.13	0.034	*
p7_ctry4:scale1	-0.18	0.05	594	-3.78	<0.001	*
mm:p7_ctry1:scale1	-0.04	0.04	594	-1.09	0.276
mm:p7_ctry2:scale1	-0.14	0.03	594	-4.82	<0.001	*
mm:p7_ctry3:scale1	0.02	0.04	594	0.53	0.599
mm:p7_ctry4:scale1	0.10	0.05	594	1.82	0.07

In both cases, we found evidence in favor of Hypothesis #2: As predicted, the main effect of porosity was significantly positive in both regression analyses (see bolded rows in tables above).

For every increase of 1 standard deviation in porosity as indexed by the Porosity (Dulin) scale, participants (on average) increased by 0.43 standard deviations in their spiritual experiences. In concrete terms, this means that for every increase of 7.63 points on the Porosity scale (range: 0-34), participants (on average) increased by 7.14 points on the Daily Spiritual Experiences scale (range: 0-70) or 7.51 points on the Spiritual Events scale (range: 0-80).

For every increase of 1 standard deviation in porosity as indexed by the Martha Story scale, participants (on average) increased by 0.32 standard deviations in their spiritual experiences. In concrete terms, this means that for every increase of 11.66 points on the Martha Story scale (range: 0-54), participants (on average) increased by 5.38 points on the Daily Spiritual Experiences scale (range: 0-70) or 5.66 points on the Spiritual Events scale (range: 0-80).

Hypothesis #3

Hypothesis #3: The positive relationship between absorption and experience is statistically significant in each field site individually.

Preregistered analysis: “For each of our five field sites, we will standardize scores on each measure (considering only participants from that field site), and then conduct the same analyses specified above for Hypothesis #1 to assess the relationship[] between absorption and spiritual experience […] in each field site individually. For Hypothesis #3, our models will take the following form: experience ~ absorption * scale + (1 | subject). […] Again, our prediction is that the main effect of absorption on experience and of each porosity measure on experience will be significantly positive (b > 0, p < 0.05).”

Here are our results:

parameter	b	standard error	df	t	p	significant
US
(Intercept)	0.00	0.09	100	0.00	1
abs	0.30	0.09	100	3.40	<0.001	*
scale1	0.00	0.04	100	0.00	1
abs:scale1	-0.06	0.04	100	-1.67	0.097
Ghana
(Intercept)	0.00	0.05	197	0.00	1
abs	0.53	0.05	197	10.48	<0.001	*
scale1	0.00	0.03	197	0.00	1
abs:scale1	0.03	0.03	197	0.79	0.433
Thailand
(Intercept)	0.00	0.07	109	0.00	1
abs	0.40	0.08	109	5.31	<0.001	*
scale1	0.00	0.04	109	0.00	1
abs:scale1	-0.03	0.05	109	-0.61	0.546
China
(Intercept)	0.00	0.09	98	0.00	1
abs	0.26	0.09	98	3.00	0.003	*
scale1	0.00	0.04	98	0.00	1
abs:scale1	-0.02	0.04	98	-0.41	0.684
Vanuatu
(Intercept)	0.00	0.08	90	0.00	1
abs	0.15	0.08	90	1.74	0.085
scale1	0.00	0.06	90	0.00	1
abs:scale1	0.01	0.06	90	0.14	0.892

We found evidence in favor of Hypothesis #3 in all sites except Vanuatu: As predicted, the main effect of absorption was significantly positive in the US, Ghana, Thailand, and China—but this was not the case in Vanuatu (although some would consider the effect “marginally significant”; see bolded rows in table above).

(See other relevant plots under “Hypothesis #1,” above.)

Hypothesis #4

Hypothesis #4: The positive relationship between absorption and experience is statistically significant in each field site individually.

Preregistered analysis: “For each of our five field sites, we will standardize scores on each measure (considering only participants from that field site), and then conduct the same analyses specified above for Hypothesis #2 to assess the relationship[…] between porosity and experience in each field site individually. […] For Hypothesis #4, our models will take the following form: experience ~ porosity * scale + (1 | subject). […] Again, our prediction is that the main effect of absorption on experience and of each porosity measure on experience will be significantly positive (b > 0, p < 0.05).”

Here are our results, using the Porosity scale (Dulin):

parameter	b	standard error	df	t	p	significant
US
(Intercept)	0.00	0.07	100	0.00	1
por	0.54	0.07	100	7.30	<0.001	*
scale1	0.00	0.04	100	0.00	1
por:scale1	-0.02	0.04	100	-0.49	0.628
Ghana
(Intercept)	0.00	0.05	197	0.00	1
por	0.57	0.05	197	11.72	<0.001	*
scale1	0.00	0.03	197	0.00	1
por:scale1	-0.05	0.03	197	-1.55	0.122
Thailand
(Intercept)	0.00	0.07	109	0.00	1
por	0.41	0.07	109	5.52	<0.001	*
scale1	0.00	0.05	109	0.00	1
por:scale1	0.00	0.05	109	0.10	0.917
China
(Intercept)	0.00	0.07	98	0.00	1
por	0.54	0.07	98	7.34	<0.001	*
scale1	0.00	0.04	98	0.00	1
por:scale1	0.05	0.04	98	1.14	0.257
Vanuatu
(Intercept)	0.00	0.08	90	0.00	1
por	0.19	0.08	90	2.25	0.027	*
scale1	0.00	0.06	90	0.00	1
por:scale1	0.11	0.06	90	1.83	0.07

Using the Porosity scale, We found evidence in favor of Hypothesis #4 in all sites: As predicted, the main effect of porosity was significantly positive in the US, Ghana, Thailand, China, and Vanuatu (see bolded rows in table above).

And here are our results, using the Martha Story scale:

parameter	b	standard error	df	t	p	significant
US
(Intercept)	0.00	0.08	100	0.00	1
mm	0.49	0.08	100	6.36	<0.001	*
scale1	0.00	0.04	100	0.00	1
mm:scale1	-0.11	0.04	100	-2.93	0.004	*
Ghana
(Intercept)	0.00	0.05	197	0.00	1
mm	0.47	0.05	197	8.76	<0.001	*
scale1	0.00	0.03	197	0.00	1
mm:scale1	-0.12	0.03	197	-3.59	<0.001	*
Thailand
(Intercept)	0.00	0.08	109	0.00	1
mm	0.33	0.08	109	4.17	<0.001	*
scale1	0.00	0.04	109	0.00	1
mm:scale1	-0.04	0.05	109	-0.83	0.411
China
(Intercept)	0.00	0.08	98	0.00	1
mm	0.38	0.08	98	4.68	<0.001	*
scale1	0.00	0.04	98	0.00	1
mm:scale1	0.00	0.04	98	0.04	0.968
Vanuatu
(Intercept)	0.00	0.09	90	0.00	1
mm	0.09	0.09	90	1.07	0.288
scale1	0.00	0.06	90	0.00	1
mm:scale1	0.03	0.06	90	0.48	0.632

Using the Martha Story scale, We found evidence in favor of Hypothesis #4 in all sites except Vanuatu: As predicted, the main effect of Martha Story was significantly positive in the US, Ghana, Thailand, and China—but this was not the case in Vanuatu (see bolded rows in table above).

(See other relevant plots under “Hypothesis #2,” above.)

Secondary exploratory analyses

“Secular” experiences

From preregistration: “First, in order to explore the question of whether affirmation of spiritual experience, porosity and absorption are simply due to the religiosity of participants, we will conduct additional analyses focusing on the two “secular” experience scales (the revised Launay Slade scale, Morrison, Wells and Nothard 2000]; and the Sheep-Goat [paranormal] scale, Thalborne & Delin, 1993). These analyses will be identical to the primary analyses outlined above, except that they will only include these two “secular experience” scales and not the two spiritual experience scales. We consider these to be exploratory analyses.”

Parallel to Hypothesis #1

parameter	b	standard error	df	t	p	significant
(Intercept)	-0.05	0.03	594	-1.63	0.104
abs	0.40	0.03	594	11.68	<0.001	*
p7_ctry1	0.04	0.06	594	0.66	0.513
p7_ctry2	0.43	0.05	594	8.31	<0.001	*
p7_ctry3	-0.09	0.06	594	-1.48	0.14
p7_ctry4	-0.27	0.07	594	-3.69	<0.001	*
scale1	0.01	0.03	594	0.53	0.594
abs:p7_ctry1	0.08	0.06	594	1.26	0.208
abs:p7_ctry2	0.06	0.05	594	1.16	0.248
abs:p7_ctry3	0.03	0.08	594	0.42	0.678
abs:p7_ctry4	0.01	0.08	594	0.12	0.903
abs:scale1	0.02	0.03	594	0.66	0.509
p7_ctry1:scale1	-0.20	0.05	594	-4.07	<0.001	*
p7_ctry2:scale1	-0.14	0.04	594	-3.40	<0.001	*
p7_ctry3:scale1	0.19	0.05	594	3.74	<0.001	*
p7_ctry4:scale1	0.07	0.06	594	1.30	0.193
abs:p7_ctry1:scale1	0.01	0.05	594	0.30	0.762
abs:p7_ctry2:scale1	-0.08	0.04	594	-1.93	0.055
abs:p7_ctry3:scale1	0.05	0.06	594	0.87	0.383
abs:p7_ctry4:scale1	0.02	0.07	594	0.24	0.812

In parallel to Hypothesis #1, the relationship between absorption and secular experience was also significantly positive (see bolded row table above).

For every increase of 1 standard deviation in absorption, participants (on average) increased by 0.4 standard deviations in their secular experiences. In concrete terms, this means that for every increase of 7.24 points on the Absorption scale (range: 0-34), participants (on average) increased by 1.52 points on the Uncommon Events scale (range: 0-18) or 0.8 points on the Extrasensory scale (range: 0-7).

This was actually a (numerically) stronger effect than what we documented for spiritual experiences (b = 0.4 vs. b = 0.27). We did not provide a prediction about whether or not this would be the case in the preregistration.

Parallel to Hypothesis #2

Here are our results, using the Porosity scale (Dulin):

parameter	b	standard error	df	t	p	significant
(Intercept)	0.06	0.04	594	1.57	0.117
por	0.51	0.04	594	12.48	<0.001	*
p7_ctry1	0.22	0.07	594	3.12	0.002	*
p7_ctry2	-0.08	0.06	594	-1.34	0.18
p7_ctry3	-0.17	0.07	594	-2.40	0.017	*
p7_ctry4	0.42	0.10	594	4.41	<0.001	*
scale1	0.15	0.03	594	4.50	<0.001	*
por:p7_ctry1	0.15	0.07	594	2.10	0.036	*
por:p7_ctry2	0.01	0.06	594	0.12	0.901
por:p7_ctry3	-0.14	0.08	594	-1.70	0.089
por:p7_ctry4	0.20	0.09	594	2.23	0.026	*
por:scale1	0.15	0.03	594	4.53	<0.001	*
p7_ctry1:scale1	-0.29	0.05	594	-5.36	<0.001	*
p7_ctry2:scale1	-0.21	0.05	594	-4.63	<0.001	*
p7_ctry3:scale1	0.13	0.06	594	2.43	0.015	*
p7_ctry4:scale1	0.39	0.07	594	5.20	<0.001	*
por:p7_ctry1:scale1	-0.04	0.05	594	-0.71	0.481
por:p7_ctry2:scale1	-0.24	0.04	594	-5.31	<0.001	*
por:p7_ctry3:scale1	0.09	0.06	594	1.44	0.15
por:p7_ctry4:scale1	0.36	0.07	594	5.10	<0.001	*

And here are our results, using the Martha Story scale:

parameter	b	standard error	df	t	p	significant
(Intercept)	0.06	0.04	594	1.55	0.121
mm	0.39	0.04	594	9.55	<0.001	*
p7_ctry1	0.30	0.07	594	4.29	<0.001	*
p7_ctry2	-0.12	0.06	594	-2.15	0.032	*
p7_ctry3	-0.24	0.06	594	-3.77	<0.001	*
p7_ctry4	0.18	0.09	594	2.10	0.036	*
scale1	0.10	0.03	594	3.55	<0.001	*
mm:p7_ctry1	0.23	0.07	594	3.33	<0.001	*
mm:p7_ctry2	0.11	0.05	594	2.08	0.038	*
mm:p7_ctry3	-0.14	0.08	594	-1.80	0.073
mm:p7_ctry4	0.18	0.10	594	1.78	0.076
mm:scale1	0.07	0.03	594	2.25	0.025	*
p7_ctry1:scale1	-0.25	0.05	594	-4.68	<0.001	*
p7_ctry2:scale1	-0.11	0.04	594	-2.60	0.009	*
p7_ctry3:scale1	0.12	0.05	594	2.45	0.015	*
p7_ctry4:scale1	0.23	0.07	594	3.41	<0.001	*
mm:p7_ctry1:scale1	-0.01	0.05	594	-0.26	0.794
mm:p7_ctry2:scale1	-0.23	0.04	594	-5.49	<0.001	*
mm:p7_ctry3:scale1	0.07	0.06	594	1.11	0.266
mm:p7_ctry4:scale1	0.29	0.08	594	3.60	<0.001	*

In parallel to Hypothesis #2, in both cases, the relationship between porosity and secular experience was significantly positive (see bolded rows in tables above).

For every increase of 1 standard deviation in porosity as indexed by the Porosity (Dulin) scale, participants (on average) increased by 0.51 standard deviations in their secular experiences. In concrete terms, this means that for every increase of 7.63 points on the Porosity scale (range: 0-34), participants (on average) increased by 1.93 points on the Uncommon Events scale (range: 0-18) or 1.01 points on the Extrasensory scale (range: 0-7).

For every increase of 1 standard deviation in porosity as indexed by the Martha Story scale, participants (on average) increased by 0.39 standard deviations in their secular experiences. In concrete terms, this means that for every increase of 11.66 points on the Martha Story scale (range: 0-54), participants (on average) increased by 1.46 points on the Uncommon Events scale (range: 0-18) or 0.76 points on the Extrasensory scale (range: 0-7).

These were actually (numerically) stronger effects than what we documented for spiritual experiences (Porosity: b = 0.51 vs. b = 0.43; Martha Story: b = 0.39 vs. b = 0.32). We did not provide a prediction about whether or not this would be the case in the preregistration.

Parallel to Hypothesis #3

Here are our results:

parameter	b	standard error	df	t	p	significant
US
(Intercept)	0.00	0.07	100	0.00	1
abs	0.46	0.07	100	6.50	<0.001	*
scale1	0.00	0.05	100	0.00	1
abs:scale1	0.00	0.05	100	0.09	0.93
Ghana
(Intercept)	0.00	0.05	197	0.00	1
abs	0.46	0.05	197	9.05	<0.001	*
scale1	0.00	0.04	197	0.00	1
abs:scale1	0.01	0.04	197	0.16	0.873
Thailand
(Intercept)	0.00	0.07	109	0.00	1
abs	0.41	0.07	109	6.03	<0.001	*
scale1	0.00	0.06	109	0.00	1
abs:scale1	0.02	0.06	109	0.40	0.693
China
(Intercept)	0.00	0.07	98	0.00	1
abs	0.36	0.07	98	5.12	<0.001	*
scale1	0.00	0.06	98	0.00	1
abs:scale1	-0.07	0.06	98	-1.13	0.263
Vanuatu
(Intercept)	0.00	0.08	90	0.00	1
abs	0.27	0.08	90	3.37	0.001	*
scale1	0.00	0.06	90	0.00	1
abs:scale1	-0.02	0.06	90	-0.37	0.709

In parallel to Hypothesis #3, the relationship between absorption and secular experience was significantly positive in all sites, including Vanuatu (where the relationship with spiritual experience was not significantly positive; see bolded rows in table above). We did not provide a prediction about whether or not this would be the case in the preregistration.

(See other relevant plots under “Parallel to Hypothesis #1,” above.)

Parallel to Hypothesis #4

parameter	b	standard error	df	t	p	significant
US
(Intercept)	0.00	0.07	100	0.00	1
por	0.53	0.07	100	8.03	<0.001	*
scale1	0.00	0.05	100	0.00	1
por:scale1	0.06	0.05	100	1.07	0.286
Ghana
(Intercept)	0.00	0.05	197	0.00	1
por	0.46	0.05	197	9.07	<0.001	*
scale1	0.00	0.04	197	0.00	1
por:scale1	-0.02	0.04	197	-0.42	0.674
Thailand
(Intercept)	0.00	0.07	109	0.00	1
por	0.32	0.07	109	4.37	<0.001	*
scale1	0.00	0.05	109	0.00	1
por:scale1	0.18	0.05	109	3.37	0.001	*
China
(Intercept)	0.00	0.06	98	0.00	1
por	0.46	0.06	98	7.13	<0.001	*
scale1	0.00	0.06	98	0.00	1
por:scale1	0.25	0.06	98	4.31	<0.001	*
Vanuatu
(Intercept)	0.00	0.08	90	0.00	1
por	0.23	0.08	90	2.80	0.006	*
scale1	0.00	0.06	90	0.00	1
por:scale1	-0.05	0.06	90	-0.87	0.384

In parallel to Hypothesis #4, using the Porosity (Dulin) scale, the relationship between porosity and secular experience was significantly positive in all sites, including Vanuatu (where the relationship with spiritual experience was not significantly positive; see bolded rows in table above). We did not provide a prediction about whether or not this would be the case in the preregistration.

And here are our results, using the Martha Story scale:

parameter	b	standard error	df	t	p	significant
US
(Intercept)	0.00	0.07	100	0.00	1
mm	0.52	0.07	100	7.82	<0.001	*
scale1	0.00	0.05	100	0.00	1
mm:scale1	0.02	0.05	100	0.35	0.726
Ghana
(Intercept)	0.00	0.05	197	0.00	1
mm	0.47	0.05	197	9.57	<0.001	*
scale1	0.00	0.04	197	0.00	1
mm:scale1	-0.09	0.04	197	-2.41	0.017	*
Thailand
(Intercept)	0.00	0.07	109	0.00	1
mm	0.24	0.07	109	3.16	0.002	*
scale1	0.00	0.05	109	0.00	1
mm:scale1	0.11	0.05	109	2.02	0.046	*
China
(Intercept)	0.00	0.07	98	0.00	1
mm	0.33	0.07	98	4.58	<0.001	*
scale1	0.00	0.06	98	0.00	1
mm:scale1	0.14	0.06	98	2.30	0.024	*
Vanuatu
(Intercept)	0.00	0.08	90	0.00	1
mm	0.01	0.08	90	0.06	0.95
scale1	0.00	0.06	90	0.00	1
mm:scale1	-0.03	0.06	90	-0.47	0.636

In parallel to Hypothesis #4, using the Martha Story scale, the relationship between porosity and secular experience was significantly positive in all sites except Vanuatu (where the relationship with spiritual experience was also not significantly positive; see bolded rows in table above). We did not provide a prediction about whether or not this would be the case in the preregistration.

(See other relevant plots under “Parallel to Hypothesis #2,” above.)

“Control” scales (Hypothesis #5)

From preregistration: “Second, in order to examine whether the hypothesized relationships between absorption and experience and between porosity and experience can be explained by a lower-level processes—namely, individual or cross-cultural differences in an overall tendency to agree with the questions posed by these scales (i.e., a “yes” bias), we will also examine the relationships between the two “control” scales (the Sense of Control scale, Lachman & Weaver, 1998; and the Need for Cognition scale, Cacioppo, Petty, & Kao, 1993) and our measures of absorption, porosity, and experience. This provides something of a negative control for our primary hypotheses. Our prediction is that any relationships with these control measures will be weaker than the relationships among absorption, porosity, and experience. To assess this, we will standardize scores on each measure, collapsing across field sites, and then conduct a mixed-effects linear regression, using scores on absorption, porosity, and control scales to predict reports of spiritual experiences. Using the “stats” package for R, we will specify this model as follows: experience ~ score * predictor * site + (1 | subject), where “site” is a categorical variable comparing the five field sites (effect-coded), and—critically—“predictor” is a categorical variable with the following orthogonal contrasts: (1) Scales of interest (absorption and the two porosity measures) vs. control scales; (2) absorption vs. porosity measures; (3) porosity as indexed by the Dulin scale vs. porosity as indexed by the Mental Causation scale; (4) Need for Cognition vs. Need for Control (Mastery Subscale). Because of the unusual and complex formulation of this model—considering each of these 5 “predictors” as repeated observations of the variable “score”—we will conduct separate models for the two spiritual experience scales (Daily Spiritual Experiences and Spiritual Events), rather than incorporating both spiritual experiences measures into the same model (as in Hypotheses #1-4). As in Hypotheses #1-4, if we have trouble fitting these models or there are other signs of model fit problems, we will consider the following alternative model: experience ~ score * predictor + (1 | site/subject). We consider these to be confirmatory tests, and in each case our prediction is that the two-way interaction between “score” and the first orthogonal contrast on “predictor” (comparing the predictive power of our scales of interest vs. our two control scales) will be significantly positive (b > 0, t > 2, p < 0.05).”

Although we did not number this hypothesis in the preregistration, I’m going to go ahead and call it Hypothesis #5.

NOTE: Neither of the models we specified in the preregistration were able to run. I think this is due to an error in my thinking (about which of our measures were “repeated measures”). The models reported here are very close to what we preregistered.

Here are our results using the Daily Spiritual Experiences scale:

parameter	b	standard error	t	p	significant
(Intercept)	-0.06	0.02	-3.55	<0.001	*
score	0.22	0.02	11.37	<0.001	*
predictorint_ctl	0.01	0.01	1.87	0.062
predictorabs_pmm	-0.02	0.01	-1.06	0.289
predictorpor_mm	0.00	0.03	0.05	0.958
predictorhthk_wob	0.02	0.03	0.65	0.517
p7_ctry1	-0.30	0.03	-9.18	<0.001	*
p7_ctry2	0.53	0.03	20.42	<0.001	*
p7_ctry3	-0.05	0.04	-1.44	0.15
p7_ctry4	-0.90	0.04	-24.01	<0.001	*
score:predictorint_ctl	0.06	0.01	7.24	<0.001	*
score:predictorabs_pmm	-0.04	0.02	-2.31	0.021	*
score:predictorpor_mm	0.08	0.03	2.68	0.007	*
score:predictorhthk_wob	0.05	0.03	1.51	0.131
score:p7_ctry1	0.08	0.03	2.43	0.015	*
score:p7_ctry2	0.07	0.03	2.45	0.014	*
score:p7_ctry3	-0.01	0.04	-0.19	0.849
score:p7_ctry4	-0.01	0.04	-0.19	0.852
predictorint_ctl:p7_ctry1	0.03	0.01	2.34	0.019	*
predictorabs_pmm:p7_ctry1	-0.06	0.03	-2.13	0.033	*
predictorpor_mm:p7_ctry1	0.00	0.05	0.02	0.982
predictorhthk_wob:p7_ctry1	0.03	0.05	0.48	0.633
predictorint_ctl:p7_ctry2	-0.03	0.01	-2.47	0.014	*
predictorabs_pmm:p7_ctry2	0.13	0.02	5.59	<0.001	*
predictorpor_mm:p7_ctry2	-0.02	0.04	-0.52	0.6
predictorhthk_wob:p7_ctry2	-0.01	0.04	-0.31	0.754
predictorint_ctl:p7_ctry3	0.00	0.02	0.28	0.781
predictorabs_pmm:p7_ctry3	0.01	0.03	0.23	0.822
predictorpor_mm:p7_ctry3	0.04	0.05	0.86	0.392
predictorhthk_wob:p7_ctry3	0.02	0.07	0.23	0.819
predictorint_ctl:p7_ctry4	0.02	0.01	1.63	0.104
predictorabs_pmm:p7_ctry4	-0.13	0.03	-3.78	<0.001	*
predictorpor_mm:p7_ctry4	0.10	0.07	1.37	0.169
predictorhthk_wob:p7_ctry4	-0.04	0.05	-0.72	0.469
score:predictorint_ctl:p7_ctry1	-0.01	0.01	-0.67	0.505
score:predictorabs_pmm:p7_ctry1	-0.05	0.03	-1.95	0.052
score:predictorpor_mm:p7_ctry1	0.00	0.05	0.07	0.944
score:predictorhthk_wob:p7_ctry1	-0.09	0.05	-1.87	0.061
score:predictorint_ctl:p7_ctry2	0.00	0.01	0.22	0.825
score:predictorabs_pmm:p7_ctry2	0.06	0.02	2.71	0.007	*
score:predictorpor_mm:p7_ctry2	0.01	0.04	0.13	0.893
score:predictorhthk_wob:p7_ctry2	0.02	0.04	0.41	0.678
score:predictorint_ctl:p7_ctry3	0.00	0.02	-0.11	0.915
score:predictorabs_pmm:p7_ctry3	0.04	0.03	1.24	0.214
score:predictorpor_mm:p7_ctry3	-0.01	0.06	-0.17	0.869
score:predictorhthk_wob:p7_ctry3	0.01	0.06	0.17	0.866
score:predictorint_ctl:p7_ctry4	0.02	0.02	1.35	0.177
score:predictorabs_pmm:p7_ctry4	-0.05	0.04	-1.35	0.178
score:predictorpor_mm:p7_ctry4	-0.02	0.07	-0.25	0.803
score:predictorhthk_wob:p7_ctry4	0.02	0.06	0.33	0.74

And here are our results using the Spiritual Events scale:

parameter	b	standard error	t	p	significant
(Intercept)	-0.07	0.02	-3.59	<0.001	*
score	0.21	0.02	10.32	<0.001	*
predictorint_ctl	0.00	0.01	-0.09	0.931
predictorabs_pmm	0.00	0.02	-0.12	0.906
predictorpor_mm	0.00	0.03	0.10	0.921
predictorhthk_wob	-0.02	0.03	-0.69	0.489
p7_ctry1	-0.17	0.04	-4.65	<0.001	*
p7_ctry2	0.57	0.03	20.27	<0.001	*
p7_ctry3	-0.22	0.04	-5.57	<0.001	*
p7_ctry4	-0.56	0.04	-13.82	<0.001	*
score:predictorint_ctl	0.07	0.01	8.34	<0.001	*
score:predictorabs_pmm	-0.03	0.02	-1.84	0.066
score:predictorpor_mm	0.02	0.03	0.69	0.49
score:predictorhthk_wob	0.12	0.03	3.50	<0.001	*
score:p7_ctry1	0.07	0.03	2.13	0.033	*
score:p7_ctry2	0.13	0.03	4.54	<0.001	*
score:p7_ctry3	0.03	0.04	0.65	0.514
score:p7_ctry4	-0.09	0.04	-2.06	0.039	*
predictorint_ctl:p7_ctry1	0.03	0.01	2.19	0.028	*
predictorabs_pmm:p7_ctry1	-0.09	0.03	-2.92	0.003	*
predictorpor_mm:p7_ctry1	-0.05	0.06	-0.80	0.421
predictorhthk_wob:p7_ctry1	0.00	0.06	0.01	0.991
predictorint_ctl:p7_ctry2	-0.04	0.01	-3.87	<0.001	*
predictorabs_pmm:p7_ctry2	0.18	0.02	7.30	<0.001	*
predictorpor_mm:p7_ctry2	0.01	0.05	0.27	0.79
predictorhthk_wob:p7_ctry2	-0.04	0.04	-0.85	0.396
predictorint_ctl:p7_ctry3	0.00	0.02	0.13	0.897
predictorabs_pmm:p7_ctry3	-0.01	0.03	-0.17	0.867
predictorpor_mm:p7_ctry3	0.03	0.06	0.52	0.606
predictorhthk_wob:p7_ctry3	-0.02	0.08	-0.21	0.832
predictorint_ctl:p7_ctry4	0.02	0.02	1.24	0.214
predictorabs_pmm:p7_ctry4	-0.09	0.04	-2.53	0.012	*
predictorpor_mm:p7_ctry4	0.04	0.07	0.52	0.601
predictorhthk_wob:p7_ctry4	-0.01	0.05	-0.10	0.92
score:predictorint_ctl:p7_ctry1	0.02	0.01	1.28	0.201
score:predictorabs_pmm:p7_ctry1	-0.05	0.03	-1.80	0.071
score:predictorpor_mm:p7_ctry1	-0.03	0.06	-0.50	0.615
score:predictorhthk_wob:p7_ctry1	-0.08	0.05	-1.60	0.109
score:predictorint_ctl:p7_ctry2	0.08	0.01	7.08	<0.001	*
score:predictorabs_pmm:p7_ctry2	-0.02	0.02	-0.76	0.448
score:predictorpor_mm:p7_ctry2	0.03	0.05	0.69	0.489
score:predictorhthk_wob:p7_ctry2	0.19	0.05	3.98	<0.001	*
score:predictorint_ctl:p7_ctry3	-0.03	0.02	-1.71	0.088
score:predictorabs_pmm:p7_ctry3	0.04	0.04	1.18	0.238
score:predictorpor_mm:p7_ctry3	0.01	0.06	0.17	0.865
score:predictorhthk_wob:p7_ctry3	-0.11	0.07	-1.63	0.102
score:predictorint_ctl:p7_ctry4	-0.01	0.02	-0.60	0.549
score:predictorabs_pmm:p7_ctry4	-0.01	0.04	-0.30	0.767
score:predictorpor_mm:p7_ctry4	-0.01	0.08	-0.09	0.925
score:predictorhthk_wob:p7_ctry4	-0.03	0.06	-0.52	0.604

In both cases, we found evidence in favor of Hypothesis #5: As predicted, the interaction between “score” and the first contrast on the “predictor” parameter was significantly positive (see bolded row in tables above), whether indexed spiritual experiences using the DSE or SE scales. In other words, the “predictors of interest” (absorption and the two porosity measures) were better predictors of spiritual experience than our two “control” predictors.

Overall scores

Histograms for each scale

Absorption

Daily spiritual experiences

Spiritual events

Ways of being

Uncommon events

Extrasensory

Hard thinking

Porosity

Martha story

Demographics

[1] "Participants who passed all attention checks"

[1] "All participants"

[1] "Participants who passed all attention checks"

[1] "All participants"

Important to note that a very substantial majority of participants in Thailand were female, and a slight majority of participants in Vanuatu were female.

[1] "Participants who passed all attention checks"

[1] "All participants"